Mapping Provincial Priorities for Educational Facility Equalization in Indonesia Using Multi-Level Clustering and Village-Based Availability

Abstract

Equitable distribution of educational facilities is crucial for development planning, as regional disparities in facility availability can constrain access to education across levels. This study identifies priority areas for school-facility equalization in Indonesia using multi-level clustering on 2024 data covering 38 provinces and the number of villages/urban wards with facilities by education level. After data cleaning, feature engineering (totals and level shares), log1p transformation, and standardization, the optimal cluster structure is selected using Elbow (inertia) and Silhouette criteria. Level-1 K-Means yields K=3 (Silhouette ≈ 0.493), classifying provinces into High Priority (15 provinces; mean total ≈ 999), Medium Priority (19 provinces; mean total ≈ 3,752), and Low Priority (4 provinces; mean total ≈ 14,362). Level-2 analysis within the high-priority group is most stable at K2=2 (Silhouette ≈ 0.51), distinguishing provinces dominated by SD–SMP shares with relatively low SMA/SMK/HE coverage from those with a more balanced composition and higher tertiary share. Overall, the framework provides a data-driven priority map and level-specific need profiles to support more targeted educational facility equalization planning.